Most quality departments I have worked in drew their boundary at the end of the production line. Once the product passed final inspection and entered the packing area, it became "logistics" — somebody else's problem. That boundary is a fiction. I have seen more customer complaints originate in the warehouse and during transport than in several assembly operations combined.
The part left the line conforming and arrived at the customer non-conforming, and nobody in the quality organisation owned the process that changed it. That is the definition of an uncontrolled process, and it deserves the same treatment as any other: a defined flow, identified failure modes, installed controls, audits and measured outputs.
Four failure categories dominate: stock rotation, physical damage, shelf-life expiry, and picking errors. Each creates defects that no amount of good process control upstream can prevent. Each is detectable and controllable — if someone defines it.
The Warehouse Performs Transformations, Whether You Acknowledge Them or Not
Ask quality managers to describe their warehouse process and the usual answer is a shrug. There is no process flow, no failure mode analysis, no defined controls. Yet the warehouse performs transformations: it changes product location, quantity, packaging condition, batch identity and time-in-storage. Every one of those transformations can go wrong, and every one is controllable if you bother to define it.
Start by mapping what actually happens to the goods: receive, verify identity, put away, store, replenish pick faces, pick, pack, despatch, load, transport, unload. For each step ask three questions: what characteristic could deteriorate here, what causes it, and how would I know? Time degrades elastomers, adhesives, sealants and certain polymers. Stacking deforms packaging and sometimes the product. Handling introduces impacts, scratches and contamination. Multi-line picking introduces the wrong part in the right box.
The output of this exercise is not a document to file. It is the basis for deciding where controls must sit. In my experience the highest-value controls are mundane: a functioning stock rotation system, racking discipline, environmental monitoring where the product demands it, and a pick verification step that actually catches errors. None require sophisticated technology. All require someone to own the process and audit it.

FIFO: The Control Everyone Claims and Nobody Verifies
First-in, first-out is the most commonly asserted and least commonly audited control in the warehouse. Ask any site whether they run FIFO and the answer is yes. Walk the racking with a batch-number scanner and you will frequently find the opposite: partial pallets of old stock at the back, new receipts in front of them, and a pick face replenished from whatever the forklift driver found first.
FIFO fails silently. Nothing looks wrong until a customer receives a batch a year older than the one you shipped the week before, or a shelf-life item arrives at their dock with a quarter of its usable life remaining. The failure modes are predictable. Physical put-away discipline is first: if the storage location system allows new stock in front of old without a system transaction enforcing it, human convenience will defeat the policy every time.
Location design is second: flow-through racking and single-deep pick faces make rotation automatic, while block stacking and drive-in racking make it entirely dependent on operator behaviour. System logic is third — ERP and WMS configurations where the removal strategy is set to "FIFO" on paper but batch-managed material is actually removed by whoever picks it.
Verification separates real FIFO from decorative FIFO. Audit it as you would any control: sample storage locations, compare the batch or date sequence physically present against the system, and record exceptions. For short-life material, monthly is a reasonable starting frequency. Measure exceptions per locations audited, put the number on the warehouse scorecard, and treat a deteriorating trend as seriously as a rising defect rate on the line — because that is precisely what it is.
Damage: The Defect Mode That Hides in Transit Reports
Physical damage between the line and the customer takes three distinct forms, each needing a different control. Storage damage comes from stacking, compression, racking contact and environmental exposure. Handling damage comes from forklift impacts, drops and dragging. Transport damage comes from load movement, inadequate restraint, vibration and moisture. Treating them as one category called "transport damage" guarantees you will never find the true cause, because the analysis stops at the loading dock.
Damage Cause Tracing
- 01Damage foundAt picking, despatch check or customer report
- 02Classify originStorage, handling or transport — not one bucket
- 03Inspect the interfaceRacking standards, forklift practice or load restraint
- 04Verify at the dockObserve loading or customer unloading directly
- 05Fix by categoryStack-height rule, handling standard or restraint method
Storage damage prevention is largely racking standards and stack-height rules. Define maximum stacking heights per pack format based on the packaging qualification, not habit. If the packaging was validated two pallets high, storing it four high is an unauthorised process change — no different from changing a weld parameter without approval. Check compression marks on cartons at the bottom of a stack: crushing means the pack format is overloaded or the pallet overstretched. Rack inspection itself belongs in the quality system whether the quality department thinks so or not.
Handling and transport damage require you to look at the interfaces. Forklift tines through the bottom carton are a training and handling-standard issue, and the evidence is visible on the pallet long before a customer complains. For transport, restraint matters as much as packaging: lateral movement in an incompletely filled trailer destroys pack integrity regardless of carton design. Where customers report transit damage repeatedly, send someone to observe unloading at their dock. I have solved more transit damage problems by watching a customer's unloading method than by any amount of packaging redesign.
Measure damage found at picking, at despatch check, and reported by customers, split by cause category and pack format. The gap between internal finds and customer reports tells you whether your own inspection is working. If customers are finding what your warehouse missed, your despatch check needs attention — not your customers.
Shelf-Life Control: Time in Storage Is a Product Characteristic
For any product with a defined shelf life — elastomeric seals, adhesives, potting compounds, lubricants, coated abrasives, certain plastics — the remaining usable life at despatch is a product characteristic every bit as real as a dimensional tolerance. Yet it is rarely treated with the same rigour. The consequence: a conforming part ships non-conforming purely through elapsed time, with no process step designed to catch it.
A conforming part can become non-conforming in storage with no process step designed to catch it — unless you design one.
The control system has three elements. Identification: every shelf-life item must carry its batch and expiry or cure date in a form readable at the point of use, not buried in a system record. Visibility: the WMS or ERP must flag approaching expiry with enough lead time to quarantine, use internally where permitted, or scrap. A despatch rule: many customers, particularly in aerospace, specify a minimum remaining shelf life on receipt; without an internal rule at least as strict, you will discover the requirement via a rejection.
Cure dates deserve special mention for elastomers, because the clock runs from manufacture, not from receipt, and storage temperature changes the rate. A warehouse that runs hot in summer is quietly consuming shelf life faster than your paperwork assumes. If temperature-sensitive material sits in an uncontrolled building, either control the building or shorten your internal life limits — deliberately, with a documented basis, not by ignoring the physics.
Audit shelf-life stock as you audit FIFO: sample the storage area, verify dates are legible and system records match physical stock, and check that near-expiry material is actually flagged and segregated rather than sitting in the pick face waiting for a deadline-driven operator. Expired material mixed into usable stock is the single worst failure mode in this category, because the picking process has no way to catch it.
Picking Errors: The Wrong Part in the Right Box
A picking error delivers a defect with perfect presentation: correct packaging, correct labels on the outer, correct paperwork — wrong contents. It is the failure mode most invisible to receiving inspection and most damaging to trust, because the customer discovers it at the point of use, often on their production line. In automotive sequenced supply it can stop a line. In aerospace it can ground an aircraft.
The root causes cluster around similarity and verification. Part numbers differing by one character, components that look identical across variants, multi-line orders picked into one container, locations stocked with the wrong material by a hurried put-away. The control question is simple: does any independent check exist between the picker's hand and the sealed box? If not, the error rate is whatever the picker's accuracy happens to be that day, influenced by workload, lighting, congestion and time pressure.
Pick Verification Options
No verification
- Error rate equals picker accuracy that day
- Wrong part found by customer at point of use
- Similar variants picked interchangeably
- No data to split error types
Independent check
- Scan verification of location, batch, quantity at pick and pack
- Or a second person verifies contents before sealing
- Pick lists designed to remove ambiguity and arithmetic
- Errors split by type: part number, quantity, batch, origin
Bar-code or RFID scan verification at pick and pack — where the system confirms location, batch and quantity before allowing despatch confirmation — is the robust solution, and I would not run a multi-variant warehouse without it. Below that, a separate pack verification, where one person picks and another verifies contents against the order before sealing, catches most errors at the cost of a second touch. Pick-list design matters more than people expect: clear print, unambiguous bin locations, one order per list, and quantities that require no arithmetic.
Measure picking accuracy as errors per lines picked, using internal catches at pack check and customer reports together. Then split the errors by type: wrong part number, wrong quantity, wrong batch, wrong location origin. Each type points to a different fix. Wrong-batch errors point to put-away discipline; wrong-part errors point to similarity management and scan verification.
Putting Logistics Under the Quality System
None of this requires new theory. It requires extending the discipline you already apply on the production line to a process that currently runs on habit. Define the warehouse process, identify its failure modes, install the controls, audit them, measure the outputs, and review the results with the same seriousness you review scrap.
The customer at the receiving dock cannot tell whether the defect was machined, stored, dropped or mis-picked — and your quality system should not pretend the distinction excuses anyone. In the standards this is not exotic: ISO 9001 and AS9100 both demand control of preservation, including storage conditions and shelf-life management, and IATF 16949 auditors will ask how FIFO is verified, not asserted.
Start with three things: an unannounced FIFO and shelf-life audit of the storage areas; a damage and pick-error Pareto built from complaints and internal finds; and a walk of the despatch area to see whether any verification actually happens between pick and seal. The findings from those three exercises will fund the argument for treating logistics as a quality process far better than any policy statement could.
